Watch the video, transcript and sources
Java is the memory-safe language, which makes this week's story awkward. Calif, a security research group, says 64 letter A's followed by an equals sign make OpenJDK write 62 bytes past the end of an array. The bug is CVE-2026-114938, nicknamed Java4Shell, and OpenJDK and Oracle published fixes on Tuesday.
Why a memory-safe language can do that. Java checks memory accesses, but the runtime under it is C, C++ and hand-written assembly. When a program decodes Base64 often enough, the JIT compiler swaps in an AVX-512 assembly routine. For the 65-character input, Java reserves a 50-byte result array. The routine finishes the padded tail with one 64-byte store and a mask. Calif's pseudocode shows the length reaching zero, the padding subtracting one more, and the mask reading minus one as "every bit on". The store starts near the end of the array, and 62 of its bytes land on whatever object sits next door.
The error comes too late. Java then rejects the input with an IllegalArgumentException. The exception is correct, and the write already happened.
What Calif says it did with that. The spill is mostly a fixed pattern, so Calif aims it at the length field of a neighboring array, lets that array accept chosen bytes, and corrupts a second array that gets copied into the response, which leaks heap memory. Calif says it reached code execution in eleven products: Keycloak, Jira, Confluence, Artifactory, Jenkins and PingFederate without a login, and Kafka, Solr, Elasticsearch, Gerrit and Spinnaker after one. Those are Calif's own tests in default setups, the recipes are withheld until people patch, and Calif itself says there is no universal exploit: every product needs its own recipe and the timing is sensitive to network jitter. Calif's account of the AI effort is one hour to find the bug, two more to build the first exploit, under $200 in tokens. We have not verified any of it.
What is confirmed. OpenJDK and Oracle list the bug at CVSS 3.1 8.1; Calif scores it 9.2 on CVSS 4.0. The public November security branch of OpenJDK contains a four line change titled "Base64 optimization improvements". CISA urged immediate patching on November 17, and when we checked on November 18 the CVE was not on its Known Exploited Vulnerabilities list. The security firm Hadrian says it crashed login checks on two lab servers and reached code execution in a lab service. Red Hat, Amazon and Adoptium confirm the scope below.
Which of your servers can run the broken code. Four things have to line up. First, an x86-64 chip with AVX-512 VBMI: Intel from Ice Lake on or AMD from Zen 4 on. ARM machines are not affected. Second, Java 17 or newer from OpenJDK builds (Oracle also lists its Java 8 performance build; standard Java 8 and 11 are not on this CVE). Third, a decoder the compiler has warmed up. Fourth, input an attacker can reach. Calif's check on Linux:
grep -m1 '^flags' /proc/cpuinfo | tr ' ' '\n' | grep -E '^(avx512f|avx512bw|avx512vbmi|bmi2)$'All four flags must appear for a machine to be in scope.
What to change today. Update to a fixed build: Oracle JDK 17.0.21.1 shipped on November 17, Amazon Corretto has fixed builds, Temurin builds are rolling out as testing finishes, and Red Hat is expediting. Until you can, start Java with -XX:-UseBASE64Intrinsics, which works on every chip. The other flag, -XX:UseAVX=2, is x86 only and stops ARM JVMs from starting. Rebuild container images, because the Java inside an old image stays old.
Verdict: SHIP IT. Ship the patch because it is small, public and already built by the main vendors, and the one flag works everywhere. Calif's code execution claims stay Calif's until someone else shows them, so update without panic. Calif is the same team behind PHP2Shell, which we covered on October 22.
Credits: portrait on the cover is James Gosling, Java's creator, photo by Peter Campbell, CC BY-SA 4.0 via Wikimedia Commons, cropped and adapted by The Daily Diff; he did not author, approve or react to this research.
Sources
https://calif.io/java4shell
https://openjdk.org/groups/vulnerability/advisories/2026-11-17
https://www.oracle.com/security-alerts/cspunov2026.html
https://www.oracle.com/java/technologies/javase/17-0-21-1-relnotes.html
https://github.com/openjdk/jdk/pull/33936
https://github.com/openjdk/jdk/commit/5177a70b385c16f26ede191fdac7aaba0bed0f8b
https://www.cisa.gov/news-events/alerts/2026/11/17/critical-java-vulnerability-requires-immediate-patching
https://www.cisa.gov/sites/default/files/feeds/known_exploited_vulnerabilities.json
https://access.redhat.com/security/vulnerabilities/RHSB-2026-018
https://adoptium.net/news/2026/11/temurin-november-2026-cspu-java4shell
https://github.com/corretto/corretto-25/blob/25.0.5.1.2/CHANGELOG.md
https://explore.alas.aws.amazon.com/CVE-2026-114938.html
https://hadrian.io/vulnerability-alerts/oracle-critical-security-patch-update-java-17-rce-claims
https://news.ycombinator.com/item?id=50538240
And that's the diff for today. I'm Niko from Axrisi. Merge responsibly.
YouTube · Newsletter · thedailydiff.dev · forward this to the intern who deployed on Friday.

